Modular
CLARA presents a software LEGO system to design and deploy scientific data processing applications without writing a single line of code. It is a data-in-motion platform to build streaming scientific-data analytics applications.
Resilient
CLARA system stays responsive in the face of a failure. Resilience is achieved by service replication, failure containment, isolation and delegation. Failures are contained within each service, isolating services from each other and thereby ensuring that parts of the system can fail and recover without compromising the system as a whole.
Elastic
CLARA system stays responsive under varying workload. It reacts to changes in the input data-stream rate by increasing or decreasing the resources allocated to process an input stream. CLARA implements predictive, as well as reactive scaling algorithms by providing relevant live performance measures. We achieve elasticity in a cost-effective way on commodity hardware and software platforms.
Robust
CLARA uses asynchronous message-passing to establish boundaries between services that ensure loose coupling, isolation, location transparency, and provides means to delegate errors as messages. Employing explicit message-passing enables load balancing and overall data-flow, i.e. application algorithm control and orchestration.
